Left ventricular myocard hypertrophy E85
Conditions
Interventions
Group 1: Phase 1 (retrospective phase):
Retrospektive analysis of already collected clinical data and imaging data from patients with echocardiographically documented left ventricular myocardial hyper
Sponsors
Universitätsklinikum Heidelberg, Klinik für Kardiologie, Angiologie und Pneumologie
Eligibility
Sex/Gender
All
Age
18 Years to No maximum
Inclusion criteria
Inclusion criteria: Phase 1 (retrospective phase): - Echocardiographically documented left ventricular hypertrophy (wall thickness =13 mm) - Digital availability of image data (echocardiography, ECG, MRI if applicable) - Complete echocardiography series available and us2.ai software compatibility Phase 2 (prospective phase): - Adult age (=18 years) - Patients identified in phase 1 with: - Echocardiographically documented LVH (wall thickness =13 mm) - High DL-based probability of cardiac amyloidosis (top 110 us2.ai score) - No evidence of definitive further cardiac amyloidosis diagnostics (e.g., no previous CMR, scintigraphy, biopsy, or genetic diagnostics) - Willingness to participate in the study and to undergo a cardiac MRI - Ability to give informed consent
Exclusion criteria
Exclusion criteria: Phase 1 (retrospective phase): - Age <18 years - Objection to the use of clinical data for research purposes - Missing or insufficient image quality - Lack of confirmed diagnosis Phase 2 (prospective phase): - MRI contraindications (implants not compatible with MRI, severe renal insufficiency GFR<30 ml/min/1.73m², known gadolinium allergy) - Known alternative diagnosis (e.g., genetically confirmed Fabry syndrome) - Pregnancy - Lack of written consent to participate in the study
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| Evaluation of the diagnostic accuracy (sensitivity/specificity, PPV/NPV, AUC) of deep learning-based algorithms (us2.ai, EchoNet-LVH, AI-ECG, AI-Hemodynamics) in the differential diagnosis of patients with left ventricular myocardial hypertrophy. Especially, evaluation of the diagnostic accuracy (sensitivity, specificity, AUC) of AI-based echocardiography algorithms (us2.ai, EchoNet-LVH) for differentiating between cardiac amyloidosis and HCM in patients with myocardial hypertrophy. | — |
Secondary
| Measure | Time frame |
|---|---|
| - Evaluation of the performance of multimodal AI-supported analysis compared to conventional diagnostics. - Analysis of diagnostic performance in different patient care pathways (emergency room, specialist outpatient clinic, referral by general practitioner). - Comparison of the analysis time between automated evaluations and manual assessment. - Development and validation of proprietary deep learning models for the classification of hypertrophic phenotypes based on combined image and clinical data. - Prospective evaluation of the influence of automated deep learning predictions on the selection of patients for further diagnostics. - Analysis of diagnostic efficiency and accuracy in real-world use. - Estimation of possible misclassifications and their consequences in the diagnostic pathway. | — |
Countries
Germany
Contacts
Public ContactBenjamin Meder
Universitätsklinikum Heidelberg, Klinik für Kardiologie, Angiologie und Pneumologie
Outcome results
None listed